University of Manchester - Investigation of the thermal and optical … · 2018. 1. 30. · Investigation of the thermal and optical performance of a Spatial Light Modulator with

The University of Manchester Research

Investigation of the thermal and optical performance of aSpatial Light Modulator with high average powerpicosecond laser exposure for Materials ProcessingApplicationsDOI:10.1088/1361-6463/aaa948

Document VersionAccepted author manuscript

Link to publication record in Manchester Research Explorer

Citation for published version (APA):Zhu, G., Whitehead, D., Perrie, W., Allegre, O., Volle, V., Li, Q., Tang, Y., Dawson, K., Jin, Y., Edwardson, S., Li,L., & Dearden, G. (2018). Investigation of the thermal and optical performance of a Spatial Light Modulator withhigh average power picosecond laser exposure for Materials Processing Applications. Journal of Physics D:Applied Physics. https://doi.org/10.1088/1361-6463/aaa948Published in:Journal of Physics D: Applied Physics

Citing this paperPlease note that where the full-text provided on Manchester Research Explorer is the Author Accepted Manuscriptor Proof version this may differ from the final Published version. If citing, it is advised that you check and use thepublisher's definitive version.

General rightsCopyright and moral rights for the publications made accessible in the Research Explorer are retained by theauthors and/or other copyright owners and it is a condition of accessing publications that users recognise andabide by the legal requirements associated with these rights.

Takedown policyIf you believe that this document breaches copyright please refer to the University of Manchester’s TakedownProcedures [http://man.ac.uk/04Y6Bo] or contact [email protected] providingrelevant details, so we can investigate your claim.

Download date:04. Jun. 2021

https://doi.org/10.1088/1361-6463/aaa948https://www.research.manchester.ac.uk/portal/en/publications/investigation-of-the-thermal-and-optical-performance-of-a-spatial-light-modulator-with-high-average-power-picosecond-laser-exposure-for-materials-processing-applications(fa10929b-e33d-485d-8a9e-fc8b2859dab1).html/portal/olivier.allegre.html/portal/lin.li.html/portal/lin.li.htmlhttps://www.research.manchester.ac.uk/portal/en/publications/investigation-of-the-thermal-and-optical-performance-of-a-spatial-light-modulator-with-high-average-power-picosecond-laser-exposure-for-materials-processing-applications(fa10929b-e33d-485d-8a9e-fc8b2859dab1).htmlhttps://www.research.manchester.ac.uk/portal/en/publications/investigation-of-the-thermal-and-optical-performance-of-a-spatial-light-modulator-with-high-average-power-picosecond-laser-exposure-for-materials-processing-applications(fa10929b-e33d-485d-8a9e-fc8b2859dab1).htmlhttps://doi.org/10.1088/1361-6463/aaa948

Journal of Physics D: Applied Physics

ACCEPTED MANUSCRIPT

Investigation of the thermal and optical performance of a Spatial LightModulator with high average power picosecond laser exposure forMaterials Processing ApplicationsTo cite this article before publication: Guangyu Zhu et al 2018 J. Phys. D: Appl. Phys. in press https://doi.org/10.1088/1361-6463/aaa948

Manuscript version: Accepted Manuscript

Accepted Manuscript is “the version of the article accepted for publication including all changes made as a result of the peer review process,and which may also include the addition to the article by IOP Publishing of a header, an article ID, a cover sheet and/or an ‘AcceptedManuscript’ watermark, but excluding any other editing, typesetting or other changes made by IOP Publishing and/or its licensors”

This Accepted Manuscript is © 2018 IOP Publishing Ltd.

During the embargo period (the 12 month period from the publication of the Version of Record of this article), the Accepted Manuscript is fullyprotected by copyright and cannot be reused or reposted elsewhere.As the Version of Record of this article is going to be / has been published on a subscription basis, this Accepted Manuscript is available for reuseunder a CC BY-NC-ND 3.0 licence after the 12 month embargo period.

After the embargo period, everyone is permitted to use copy and redistribute this article for non-commercial purposes only, provided that theyadhere to all the terms of the licence https://creativecommons.org/licences/by-nc-nd/3.0

Although reasonable endeavours have been taken to obtain all necessary permissions from third parties to include their copyrighted contentwithin this article, their full citation and copyright line may not be present in this Accepted Manuscript version. Before using any content from thisarticle, please refer to the Version of Record on IOPscience once published for full citation and copyright details, as permissions will likely berequired. All third party content is fully copyright protected, unless specifically stated otherwise in the figure caption in the Version of Record.

View the article online for updates and enhancements.

This content was downloaded from IP address 130.88.107.48 on 30/01/2018 at 16:03

https://doi.org/10.1088/1361-6463/aaa948https://creativecommons.org/licences/by-nc-nd/3.0https://doi.org/10.1088/1361-6463/aaa948

Investigation of the thermal and optical performance of a Spatial

Light Modulator with high average power picosecond laser

exposure for Materials Processing Applications

G. ZHU1, D. WHITEHEAD2, W. PERRIE1,* O. J. ALLEGRE2, V. OLLE3, Q. LI1, Y. TANG1, K.

DAWSON4, Y. JIN1, S. P. EDWARDSON1, L. LI2 AND G. DEARDEN1

1Laser Group, School of Engineering, Brownlow Street, University of Liverpool, L69 3GQ, UK

2Laser Processing Research Centre, School of Mechanical, Aerospace and Civil Engineering, The University of Manchester, Manchester

M13 9PL, UK 3 Hamamatsu Photonics UK Ltd., 2 Howard Court, Welwyn Garden City, Hertfordshire AL7 1BW, UK

4 Nanoinvestigation Centre at Liverpool, University of Liverpool, Liverpool, L69 3GL, UK

*[email protected]

Abstract: Spatial light modulators (SLM’s) addressed with Computer Generated Holograms (CGH’s) can

create structured light fields on demand when an incident laser beam is diffracted by a phase CGH. The power

handling limitations of these devices based on a liquid crystal layer has always been of some concern. With

careful engineering of chip thermal management, we report the detailed optical phase and temperature response

of a liquid cooled SLM exposed to picosecond laser powers up to

= 220W at 1064nm. This information is

critical for determining device performance at high laser powers. SLM chip temperature rose linearly with

incident laser exposure, increasing by only 5C at

= 220W incident power, measured with a thermal

imaging camera. Thermal response time with continuous exposure was 1-2 seconds. The optical phase response

with incident power approaches 2 radians with average power up to

= 130W, hence the operational limit, while above this power, liquid crystal thickness variations limit phase response to just over radians.

Modelling of the thermal and phase response with exposure is also presented, supporting experimental

observations well. These remarkable performance characteristics show that liquid crystal based SLM technology

is highly robust when efficiently cooled. High speed, multi-beam plasmonic surface micro-structuring at a rate R

= 8cm2s-1 is achieved on polished metal surfaces at

= 25W exposure while diffractive, multi-beam surface

ablation with average power

=100W on stainless steel is demonstrated with ablation rate of 4mm3min-1.

However, above 130W, first order diffraction efficiency drops significantly in accord with the observed

operational limit. Continuous exposure for a period of 45 minutes at a laser power of

= 160W did not result

in any detectable drop in diffraction efficiency, confirmed afterwards by the efficient parallel beam processing

at

= 100W. Hence, no permanent changes in SLM phase response characteristics have been detected. This

research work will help to accelerate the use of liquid crystal Spatial light modulators for both scientific and

ultra high throughput laser-materials micro-structuring applications.

1. Introduction

The construction of the first 50 × 50 pixel Spatial Light Modulator (SLM) based on an active silicon

backplane combined with a nematic liquid crystal as the light modulator was described in 1989 by

McKnight [1], while development of a 256 x256 array followed within 5 years [2] and commercial

development of liquid crystal SLMs has been rapid ever since by companies such as Holoeye (Germany)

Boulder Systems (USA) and Hamamatsu (Japan). SLM’s have been used for a wide range of both scientific

and industrial applications such as wavefront correction in astronomy [3], the creation of spiral laser beams

carrying Orbital Angular Momentum (OAM) [4], clarifying the relationship of OAM to Spin Angular

Momentum (SAM) [5], static and dynamic parallel beam processing with ultrafast lasers [6-9] and dynamic

polarisation control for surface micro-patterning [10]. Also, a competing technology, termed a Grating

Light Valve (GLV) is a diffractive Micro-Opto ElectroMechanical System (MOEMS) spatial light

modulator, capable of high bandwidth (kHz) modulation of light with applications in high-resolution

displays and computer-to-plate printing [10]. The average laser power handling capability of these devices

is impressive, around 60W. While in general, these devices can modulate only intensity, recent research

using a MOEMS device combined with polarising crystals demonstrated high bandwidth vector field

modulation at 5kHz [11].

Page 1 of 21 AUTHOR SUBMITTED MANUSCRIPT - JPhysD-113405.R2

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

mailto:*[email protected]://proceedings.spiedigitallibrary.org/solr/searchresults.aspx?q=Microopto+electromechanical+systems

Fixed Diffractive Optic Elements (DOEs) based on precision etched surface profiles on fused silica are

very robust, able to handle 10’s of Watts but designed and fabricated for one particular function only,

whether beam shaping or generating multi beams [12]. SLM’s on the other hand, are dynamic diffractive

optics with wide flexibility for structuring laser intensity and polarization when applied with appropriate

Computer Generated Holograms (CGH’s), calculated from inverse Fourier Transforms [13].

Power handling, (both average and peak power) with a liquid crystal based SLM has induced some

anxiety with regard to likely damage levels and thus experiments carried out at high power levels have been

limited. The first demonstration in this direction was by Beck et al [14] who cooled a commercial, Holoeye

SLM, (model LCR-2500) rated for average power

~3W maximum power, increasing this well beyond

the commercial limit to

= 14.8W with 532nm, nanosecond laser pulses. This was achieved by

mounting the chip to a large, optically flat copper cooling block. The surface profile of the chip was

measured in a stepping Michelson interferometer with the SLM placed in one arm and a flat mirror in the

reference arm with the resulting interference pattern observed on a screen by a CCD camera. By converting

the phase map (fringes) to height differences, the surface profile of the SLM display before and after

mounting onto the flat copper heat sink was determined experimentally. These measurements demonstrated

that mounting the CMOS chip altered the surface flatness significantly and by tensioning mounting screws

to the copper plate, central flatness was reduced to Ra 0.20 m rms at 532nm. Processing with 532nm radiation was then demonstrated with a 4f system and f-theta focusing lens to a fixed workpiece, imaging a

series of complex kinoforms (calculated by IFT’s) on metal coated glass (100ms exposure) and shorter,

3.2ms exposures on polyimide coated metal with 2D patterns well defined.

More recently, a pulsed fibre laser (1065nm/20kHz/1mJ) with

= 50W was integrated with a

cooled SLM (Hamamatsu X-11840-03) and galvo scanner for high speed multi- beam surface patterning

[15]. Phase CGHs calculated from IFTs defocused the zero order well away from the 10 first order spots

and clear, uniform spot letter patterns were ablated on silicon at a rate of 6.105 spots per second by the

SLM and fast galvo scanner. The same group extended their experiments to cutting of thin, 0.5mm thick

steel sheets with a CW fibre laser (1070nm) operating with

= 190W continuous exposure [16]. The

cutting quality (with reduced burr) was slightly improved with the addition of the low intensity first order

spots around the reflected high intensity zero order beam. As the fibre laser beam was randomly polarised, the diffraction efficiency was poor, reaching only 10%, estimated from the measured intensity profiles.

Very recently, Klerks and Eifel (Pulsar Photonics GmbH) demonstrated laser processing with up to

= 60W (404kHz/6ps) of laser radiation at 515nm on a cooled SLM in a fully integrated flexible beam

shaping system with active cooling of the SLM chip [17]. Phase response measurements were inferred by

exposing the SLM (Hamamatsu X13139-04, 1280x1024 pixels) simultaneously to the high average power ps

heating laser and an expanded laser pointer, which were overlapped on the SLM. The HP laser was reflected

to a beam dump while the laser pointer was directed via a 4f optical system to the input of a scanning galvo

with f-theta lens and imaged to a camera below. Using an integrated temperature chip, active feedback kept

the liquid crystal at 14 C while laser power was altered over the range

= 0-60W. After each 1W power step the grey level required for a phase change of was determined, the so called control parameter. This

was measured by applying binary gratings to 180 by 180 pixel regions and altering phase to maximise the

1 order intensities relative to the zero order intensity observed on the camera system below the f-theta lens. When

1 0I / I was maximised, the grey level (GL) for a phase change (GL =100) on a particular region was

determined and found to vary little over the full 60W exposure range. Robust parallel beam surface micro-

machining was then demonstrated on stainless steel with powers from

= 15W to 60W. At

= 60W exposure, the company logo (created using a complex kinoform) was also micro-machined cleanly but there

was also evidence of detrimental effects due to lower intensity ghost beams on the surface.

Our experience over a decade with the X-10468 (Hamamatsu) series of reflective phase only SLMs

showed that peak intensities >10GW/cm2 and average power

12W without cooling led to no detectable deterioration after years of continuous operation. These are impressive characteristics. However, the idea of

pushing these devices to much higher exposures certainly induced anxiety, based on the likely cost of chip

replacement in the event of thermal damage. The perceived limitations on robustness of the liquid crystal

layer, particularly with average power exposure has, until now, severely limited industrial uptake of this

technology, particularly for laser-materials processing.

In this paper, we first investigated the thermal response of cooled liquid crystal on Silicon (LCOS)

SLMs with picosecond laser exposure up to

= 220W. Then, the complete phase response was measured

Page 2 of 21AUTHOR SUBMITTED MANUSCRIPT - JPhysD-113405.R2

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

with incident power to ascertain device performance at high powers. Finally, with this knowledge,

diffractive multi-beam materials processing was demonstrated with up to

= 250W laser power.

2. Method and Experimental details

Two separate experimental set-ups were created for measuring chip temperature and phase response respectively

with laser exposure. For chip temperature measurements, a simple set-up, shown in the schematic of Figure 1(a)

was used. The output beam from an Edgewave, Innoslab laser, (10ps, 1064nm, 2MHz) Model PX400-3-GH,

located above an optical table was directed downwards by a 90 turning periscope which brought the beam to

required level for the SLM, a Hamamatsu X13139-03, with 1280 x 1024pixels, 12.5 m pixel size and 16 x

12mm dimension. The beam was then expanded with a telescope (x4, f1=-50mm, f2=200mm, AR coated) before

reflecting from the SLM at low angle of incidence AOI < 10 and directed to an air cooled power meter (Gentec

UP55G-500F-H12) rated for 500W. A calibrated far infrared thermal camera (model FLIR SC660, sensitive

from 8µm-15µm) was positioned around 0.5m from the SLM for capturing thermal images. It was possible also

to take real time videos with this camera.

The SLM was mounted on an engineered copper block, thermally connected to the rear of the silicon chip

with copper tubes on the side for liquid cooling of the unit. The block has internally structured cooling channels

and the circulating liquid was pumped by a Koolance (model EX2-755) liquid to air cooling unit rated for

500W. Laser repetition rates of 10kHz - 2MHz were employed by pulse selection from the oscillator. Figure

1(b) shows a photo of the mounted SLM with cooling connectors.

Figure 1. (a) Schematic of experimental set-up for high power thermal tests of cooled SLM. The raw laser beam was expanded x4, and

reflected at low angle from the SLM to the power meter. A thermal camera (FLIR SC660) was positioned to detect heat radiated from the

SLM chip and enclosure, (b) photo of cooled SLM (Hamamatsu X13139-03) on mirror mount with liquid cooling connections to the Copper

heat sink.

The phase response of the cooled SLM was investigated with the experimental set-up shown in Figure 2.

The laser output was expanded x3 and linear polarisation rotated to 45 on the SLM with a half wave plate (Altechna 2-CPW-TZ0-L2-1064). The SLM director was horizontal. A series of CGH’s with grey level GL = 0-

255(8bit) were applied, so that the SLM now behaved as a variable waveplate (introducing a phase delay

between the vertical and horizontal electric field components) hence reflected elliptical polarisations. By placing

a quarter waveplate (Altechna 2-CPW-TF0-L4-1064) with fast axis at 45 , reflected elliptical polarisation

was re-converted to linear polarisation whose rotation direction depended linearly on the applied grey level CGH. This linear polarisation was then analysed by a thin film polarizer (TFP, Altechna 2-HC45TPF-1064-

0254). The transmitted p and reflected s components were measured with air cooled power meters, Gentec

UP55G-500F-H12 (500W) and Gentec UP55N-300F-H12-DO (300W) respectively. A series of phase response


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

curves with increasing exposure were measured. The polarisation optics, which are ion beam coated have

ultralow low absorption and ultrahigh damage thresholds, quoted as Fth > 20Jcm-2 with a single ns pulse [18]

Figure 2. Optical set-up used to determine phase response of cooled SLM using polarisation modulation induced by GL phase maps

applied to the SLM. Linear polarisation, incident at θ=45° on the SLM is reflected as elliptical polarisation then converted to linear

polarization by the / 4 plate with polarisation direction dependent on the GL phase CGH applied.

Let us assume for the moment that the incident linear polarisation, after reflection from the SLM is right

circularly polarised,

1R

i

with phase difference 2

between electric field components. Using the

Jones matrix for a / 4 plate tilted at 45 the resulting polarisation after traversing the / 4 plate is given by, [19]

1 i 1 2 12 2 x

i 1 i 0 0

(1)

which is linearly polarised in the horizontal direction while the effect on left circularly polarised

1L

i

with

phase difference 2

yields vertically linearly polarised light,y

. Hence, linear polarisation is rotated though

an angle / 2 while the phase difference between the orthogonal states R

andL

,

. The rotation

angle of linear polarisation after the waveplate is thus given by/ 2

. If the linear polarisation now makes

an angle with respect to the transmission axis of the polariser, which is horizontal, the transmitted (reflected)

amplitude is given by 0E cos 0(E sin ) while the transmitted (reflected) intensities are given by

2

0I cos

2

0(I sin ) . The transmitted (reflected) powers are hence given by2

0P cos ( / 2) ,2

0(P sin ( / 2)) where 0P

is the

incident power and

is the phase applied to the SLM. This function has two maxima over 2 radian. This

simple theoretical analysis assumes that the incident laser is highly linearly polarised ( >100:1) and that the TFP

has a high contrast, which is indeed the case, stated to be p sT / T 1000 :1

with wavefront error / 8 in the

visible [16]. The blazing function of the SLM, supplied with Hamamatsu software is linear, so thatGL

.


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

3.0 Experimental Results

3.1 Thermal tests

Two X-13931-03 Hamamatsu liquid cooled devices were tested. For accurate temperature measurements, the

emissivity on the FLIR camera software was set to = 0.73 , the value relevant for Silicon (back plane) near

room temperature [20]. When setting instead to 0.93, (fused silica window), this had only a minor effect on measured absolute temperatures. Figure 3 shows the temperature response of device-01 with up to

= 140W

exposure, showing that this is highly linear with a maximum temperature rise 5.5T C and hence a thermal

response coefficient m01 = 0.041C/W, an impressive result. Laboratory ambient temperature was cool early in

the morning.

Figure 3. Measured temperature rise with reflected laser power, device-01. The thermal response coefficient is measured to be 0.041C/W.

Laser repetition rate f = 404kHz. Errors represent 1 and fit is least squares with high confidence, R = 0.99.

Chip temperature response of the liquid crystal device-02 with reflected laser power in the range

= 50-

215W is shown in Figure 4. Again, temperature rises linearly with reflected power over this range, however, the

temperature difference with

= 215W (reflected) was only 5.0T C while ambient chip temperature was

T0 = 21.8. The response of this cooled SLM is more impressive, yielding a gradient m01 = 0.026 C / W with a

cooling efficiency higher by a factor of 1.6 over device -01. As device reflectivity R = 97%, incident laser

power was therefore

= 220W. While these devices are apparently identical models, there is clearly a

difference in the thermal cooling efficiency, the source of which is not yet understood.


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

Figure.4 Measured temperature rise with reflected laser power, device-02. Laser repetition f = 1MHz. Temperature response is excellent

with a gradient m02 = 0.026C/W, superior to device -01. Errors represent 1 and fit is least squares with high confidence, R = 0.99

These highly linear characteristics imply that the effective absorption coefficient, of the chip assembly

(window, liquid crystal layer, dielectric reflective layer and silicon backplane) is constant throughout and has

only linear temperature dependence in this power range when cooled effectively.

Figure 5 shows typical thermal camera images recorded as laser exposure was increased from

= 0 to 220W,

demonstrating the excellent cooling efficiency of device -02. Vertically scattered radiation from the incident

laser beam warms the surrounding black absorbing SLM enclosure a few degrees, figures 5(b-f). Clear evidence

of the intense beam exposure appears in figure 5(e) at

= 215W near the chip centre, a few degrees higher in

temperature than the rest of the chip. The high power laser exposure combined with efficient cooling now leads

to a temperature gradient across the chip, figures 5(d-f).

Figure 5. Thermal images of cooled SLM chip with increasing reflected laser exposure (a)

= 0, (b)

= 44W, (c)

= 118W, (d)

= 172W (e)

= 215W, showing evidence of laser spot near the centre of chip, (f)

= 215W showing spot temperature T =

26.6C. Note that there was some off axis low intensity scatter from the laser beam reaching the surrounding enclosure (matt black) hence

showing a slight temperature rise above and below on this highly absorbing surface. With the scattered radiation removed with an aperture,

FLIR images showed a reduced enclosure temperature, generally cooler than the chip temperature.


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

When the laser power at

= 220W was cut off instantly, the chip temperature cooled back to ambient in

approximately 1-2 seconds, supporting the high thermal diffusion rate. With R = 97% reflectivity, almost

=

7W was efficiently removed from the SLM structure. A comparison with our previous work using an uncooled

Hamamatsu X10468-04 (532nm) at

8W exposure with a Gaussian beam showed a temperature rise T

2 or a thermal response of 0.25C/W, thus 10 times poorer than in cooled SLM, device-02.

3.2 Phase measurements

The phase modulation response of the cooled SLM’s, critical for determining SLM performance under intense

laser exposure is now presented for the experimental set-up shown above in Figure 2. Measured phase response

curves are shown in Figure 6(a-f) for total power = 7.5W, 21.8W, 37.8W, 78.8W, 91.1W and 137.5W

respectively on device-01. The fits are based on 6’th order regression polynomials. In figure 6(a), at lowest,

7.5W exposure, the expected near sinusoidal response with GL is very clear and the sum of powers

nearly constant. The error bars show one standard deviation. Where the curves cross, the polarisation incident

after reflection from the SLM is circular while at the maxima and minima, linearly polarised with a high degree

of polarisation. There is a phase change between the maxima (blue to red), minima, or crossing points. This

phase response (apart from a phase shift) follows the predicted 2cos ( / 2) ( 2sin ( / 2) ) functions, essentially

identical up to

= 38.5W, figure 6(c) while at

= 79W and

= 91W, figure7(d), figure 7(e), there is

now a deviation from the ideal phase response near the peaks but phase variation still almost 2 at GL = 225.

However, at = 132.5W exposure, the response curve has altered significantly, yielding just over a

phase change. However, note that at grey level GL =100, the reflected polarisation is still highly linear with

degree of polarisation max min max min/ P I I I I ~ 0.93 and an excellent sign that the liquid crystal layer is

not under stress and still functioning. Ofcourse, phase and wavefront errors can be expected on reflected

radiation at this laser power. The total exposure time of device -01during these phase tests was approximately

1.5 hours.


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

Figure 6. Phase response of cooled SLM -01 (a)

= 7.5W, (b)

= 21.8W, (c)

= 37.8W, (d)

= 78.8W, (e)

= 91.1W and

(f)

= 132.5W. The phase response is very satisfactory up to laser power

= 37.8W but begins to deviate slightly above 80W while a

significant deviation occurs at

= 132.5W, reaching just over a phase shift instead of the desired 2 .

The results of phase response measurements on device -02 are shown in figures 7(a-e) with power

levels over the range = 2.6W -196W. This time, the phase response curves are satisfactory up to

109W, figure 7(d) achieving a full 2 phase change coupled with a high degree of polarisation

max min max min/ 0.90 P I I I I . At

= 160W, figure 7(e) and

= 196W, figure 7(f), there is

reduced phase modulation reaching just over radians while the degree of polarisation reduces to P 0.75 and 0.64 respectively. This drop in observed degree of polarisation is likely due to the temperature gradient across the chip, altering reflected state of polarisation from centre to the beam edge. The changes in response curves infer that wavefront errors will occur during application of phase only CGHs with

incident laser powers exceeding

= 109W, altering desired relative intensities in a structured light field.

Total exposure time on device-02 during these measurements was several hours and for 45minutes at

=

160W power level. When reducing power levels after exposure at 196W, phase reponses were reproducible.


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

Figure 7. Phase response of cooled SLM -02 (a)

= 2.6W, (b),

= 26.2W, (c)

= 60.9W, (d)

= 109W, (e)

= 160W

and (f)

= 196W. The phase response is very satisfactory at laser power up to

= 109W but deviates significantly at

=

160W and above, reaching over a phase shift instead of the desired 2 .

4. Modelling of observed thermal and phase response

It is critical to explain the observed temperature and phase response from a theoretical standpoint. However, as

exact SLM structure and materials are commercially protected, approximations of these and physical

dimensions are made so that modelling may be limited by this lack of complete knowledge.

The silicon chip (1mm thick), is connected to a ceramic cooling plate (AlN with similar thermal conductivity

of 150Wm-1K-1) then thermally connected to the water cooled copper heat sink. The liquid crystal (LC)

properties, for example, the LC birefringence, its temperature response and nematic-isotropic transition range

are unknown, hence approximate values based on the literature of LC data are used to estimate the phase

response with temperature.

a) Thermal response

Assume for the moment that absorption in the window (fused silica) and LC layer are neglible. As the dielectric

coating reflectivity R = 97%, then

= 7W is absorbed in the Silicon for

= 220W incident average power.

We therefore assumed a 7W thermal heat source with a Gaussian beam distribution of 10mm 1/e2 diameter. A

thermal model using the 3D heat diffusion equation was used in COMSOL Multi-physics software which

contains the relevant physical properties such as density, heat capacity and thermal conductivity for each

material (Si, AlN and copper). Water cooling of the copper base was also introduced with flow rate 1 l/min.

Appropriate mesh sizes from 1-3mm were used for the numerical calculations.

Theoretical results are shown in figure 8. With no cooling, figure 8(a) shows the calculated temperature rise of

the Silicon top surface (red), the Si-ceramic interface (green) and ceramic-Copper interface (blue). The system,

as expected, does not come to thermal equilibrium. The surface temperature rises 2C in the first few seconds

then increases at a linear rate of 0.024C/sec. Ambient temperature was 293K.

The effect of adding water cooling of the bottom copper surface at 20C is shown in figure 8(b) demonstrating

that thermal equilibrium is reached after approximately 20seconds. The chip temperature rises rapidly in the first

second then comes to a stable temperature with a 2C rise. The expected temperatures at the interfaces are also

plotted showing that there are temperature gradients throughout, physically realistic. While the absolute

temperature rise at the silicon surface is lower than that observed by a factor of 2.5, nevertheless, this model

supports the view that absorption in Silicon is a major source of the observed temperature rise during laser

exposure of the SLM.


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

The calculated surface temperature of Silicon with absorbed power (

= 0-10W) is shown in figure 8(c),

yielding a linear rise with exposure (absorbed power) which is in agreement with experimental observation,

(figure 4). The calculated temperature distribution across the chip with exposure is shown in figure 8(d),

indicating that a thermal gradient exists across the chip, increasing with exposure in accord with the FLIR

camera observations (figure 5).

Figure 8. Expected chip temperature with

= 7W absorbed (

= 220W incident), (a) without water cooling of copper heat sink, (b)

with water cooling of copper heat sink showing that thermal equilibrium is achieved, (c) chip temperature (centre) with absorbed power, d)

Calculated temperature distribution after reaching thermal equilibrium with absorbed powers of

= 2W, 7W and 10W (incident powers of

= 63W, 220W and 315W respectively.)

The window material is likely to be fused silica which has a remarkably low absorption coefficient at 1064nm, 1064 5 1

fs 1.10 cm [21] hence we the absorbed power with incident average power

= 220W is

then variations in the optical thickness, nL will cause phase changes. After reflection from the dielectric

reflecting layer, the phase change is given by (4 / ) nL . The total change with temperature is given by,

( n) L(4 / ) L n

T T T

(1)

Assuming for the moment that the LC layer thickness is unaffected by temperature, L is constant so that the

second term in equation (1) is zero. The change of phase with temperature is then,

( n)(4 / )

T T

(2)

The variation of birefringence with temperature in LC’s is well represented by an empirical equation, [24]

0 INn n (1 T / T ) where n0 is the birefringence extrapolated to 0K, T is temperature (K), TNI is the

nematic-isotropic phase transition temperature (K) and the index is the orientational order parameter. TNI is

typically > 60C [24] while the index typically varies in the range 0.1 < < 0.3. The birefringence thus

decreases with temperature and hence phase change will drop, however, this is a slowly changing function,

except near TNI where n goes rapidly to zero. Using this form for n in equation (2) and differentiating, the phase change with temperature is,

1

0 NI NI(4 L / ) n (1 T / T ) / TT

(3)

Setting T = 293K, (20C), TNI = 333K (60C) and using a typical value of n = 0.30 at room temperature with

= 0.19, then n0 = 0.45 (0K). Using these values, equation (3) yields 0.01T

rad/K. Hence with a

temperature change T = 5C, the phase change = -0.05 radian. At 1.064nm, this corresponds only to

0.025. This very small value infers that the source of the phase response change is more likely due to

changes in the LC layer thickness L. As the LC layer thickness is given by L / 2 n , a reduction in phase

change corresponds to an effective change in L / 4 n 0.88m. This could occur through mechanical

distortion of the chip structure due to the developing temperature gradient, observed experimentally and

predicted above combined with differences in thermal expansion coefficients of the various materials.

Let us assume that the temperature gradient results in a change in the thickness of the LC layer, which is

maximum in the centre and reducing towards the edges. This varying thickness affects the phase response so

that, for a given GL, the centre produces a different phase delay than the edges. In the experiments, this phase

delay is converted into a rotation of the linear polarization (see figure 2) and thus the polarization direction will

vary from the centre to the edges. The effect of the polarizer (TFP in figure 2) is to transmit only the component

of the polarization that is parallel to its axis. For example, the transmitted power from a uniform linearly

polarized beam with polarization angle is 2

t iP P cos .

To estimate the propensity of the temperature gradient to induce a non-uniform polarization direction, we will

assume that the beam can be modelled as the sum of three concentric areas, each having a distinct uniform

polarization angle. The laser power transmitted through the TFP from the central area is 2

t1 i1P P cos whereas

the laser power transmitted from the remaining two concentric areas around the centre P t2 and Pt3 will be phase


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

shifted by varying degrees. The normalized transmitted power is the sum of all three components. An example

solution is,

2 2 2

t iP / P [1.3cos 1.3cos( / 2 / 8) 0.4(3 / 4 3 / 4) ] / 3 (4)

Figure 9 shows the resulting expected phase responses at low and high average powers. In figure 9(a) at average

power

= 26.2W, experiment is compared with a pure 2cos function showing excellent agreement with

experiment over 0 2 . Figure 9(b) shows the example phase response (above) compared with the experimental

data at high average power

= 160W. The fit of theory and experiment is in reasonable agreement and

supports the view that the variations in LC thickness across the chip due to the developing thermal gradient may

be the source of the limited phase range at high average laser powers.

Figure 9. The calculated phase response with incident average power. (a) experimental data at low average power,

= 26.2W compared

with a pure 2cos function, (b) when three regions of the thermally affected region of the chip are modelled with different phase response,

equation (4) and compared to experiment at

= 160W. This limited phase response will affect diffraction efficiency significantly.

5. Laser Micro-machining results

5.1 Parallel beam laser surface processing at

25W

The laser mode, observed on a NIR card at low powers of a few Watts was elliptical with major axis in the

horizontal. Dimensions of the raw beam were estimated to be approximately 12 x 7mm, yielding an eccentricity

e ~ 1.7. The mode at high powers remained elliptical, confirmed by observations with a hand held NIR camera

(Find-R-scope) looking at the low intensity scatter from the SLM chip and other optics. Prior to laser-micro-

machining with the cooled SLM, the TFP and / 4 plate were removed and a 4f optical system added with f1 =

300mm, f2 = 160mm (both plano-convex, AR coated) hence reducing the beam diameter by x0.53. This helped

to reduce the likelyhood of diffracted beams clipping the edges of any following 25mm diameter optics, thus

avoiding energy losses. The reflected laser output from the SLM was directed after lens L2 to the input aperture

(14mm) of a digital scanning galvo (Scanlab intelliScan 14/RTC-5 card) and focussed with a 100 mm focal

length tele-centric lens operating under Scaps GmbH scanner software. With phase CGH applied, the complex

optical field at the SLM was thus re-imaged to the input aperture of the scanning galvo. A polished stainless

steel plate, 60 x 60 x 3mm was placed on an adjustable lab jack for vertical control to bring the surface to the

focal plane. In addition, polished brass coated steel plates, 80 x 80 x 2mm were exposed for highly parallel

beam patterning. At a given power, the phase response was re-checked and Grey Level required for pecisely

2phase difference determined. This allowed one to alter the blazing function in Labview software relevant to a


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

given power level. As phase CGH’s can allow efficient use of the high power laser through diffractive splitting,

parallel beam surface micro-processing was carried out at a range of powers.

A CGH generating three spots ( 1 and zero order) on a polished stainless steel plate were created using the non-iterative lens and gratings algorithm in Labview. Figure 10(a) shows an optical image of the three ablated

parallel lines at low power with laser parameters of 10kHz,

= 2W and 250mms-1 scan speed. The lines have

approximately the same width (80-100 m ) showing that laser power was evenly distributed between the orders.

Pulse energy was therefore ~ 70 J per pulse per beam. At the end of a line scan, the decelerating beam created

an imprint of the laser intensity distribution and by fitting an ellipse to the outline in optical software, the beam

spot dimensions were estimated to be ~ 88 1 m x 31 4 m with beam ellipticity e ~ 2.8. However, most of

the energy appeared in a central lobe, approximately 32 2 m x 23 2 m giving ellipticity of the central

region as e ~ 1.4. Averaging these yields e ~ 2.1, a more realistic estimate and closer to the ~1.7 estimated using

the NIR card. Laser beam quality factors were estimated to be 2XM ~ 4.3 and

2

YM ~ 2.3 respectively from the

observed surface ablation dimensions. Single pulse fluence was therefore around F = 2.8Jcm-2 and pulse

overlap N ~25 per spot along the smaller axis.

Figure 10(b) shows the 1 (overlapped) and zero orders machined at 1ms-1 with

= 25W/404kHz, hence with pulse energy E ~21 J and fluence F ~ 0.85Jcm-2. As offset hatch spacing was 460 m , this

overlapped the st1 orders, yielding an asymmetric pattern. Figure 10(c) shows an SEM image of the asymmetric

pattern.

Figure 10. (a) Optical micrograph of three parallel lines (st1 order, 0th order middle) machined at fluence F = 2.8Jcm-2 on stainless steel

with

= 2W, 10kHz and scan speed 250ms-1, separated by 250 m with line widths ~90 m , (b) optical image of large area 3 spot

processing with

= 25W/404kHz exposure with 1ms-1 scan speed and offset hatch spacing of 460 m , overlapping the st1 orders hence

yielding an asymmetric pattern. Pulse energy E ~21 J and fluence F ~ 0.85Jcm-2, (c) SEM image of the stainless steel surface in (b) with

= 25W exposure clearly showing the overlapping st1 orders and zero order in between.

When linearly polarised ultrashort pulses are scanned on a surface with fluence near the ablation threshold,

Laser Induced Periodic Surface Structures (LIPSS) can result due to the interference of a surface scattered wave

with the incoming radiation [25,26] This interference modulates the light intensity of the radiation on the

surface, resulting in periodic ablation with pitch close to the wavelength of the exciting laser. Ultrashort laser

pulses with 10ps pulselengths are ideal for encoding diffractive, periodic surface structures since heat


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

diffusion and melting are minimised during ablation [27] White light is easily diffracted by these periodic

surface undulations. Figure 11(a) shows the result of large area parallel beam surface ablation (3 spots) at

= 25W average power, 404kHz repetition rate and 3ms-1 scan speed. The colour differences are due to the slight

change in viewing angles of the camera lens. Pulse energy was hence around 21 J and average power

~

8.3W per beam. With focused beam area of ~2.5. 10-5cm2, the single pulse fluence is approximately F ~

0.85Jcm-2. As the hatch spacing was 260 m , the whole surface was patterned and effective patterning rate R ~

8cm2s-1. Figure 11(b) clearly shows the periodic surface micro-structure (SEM) parallel to the scan direction

while figure 11(c) a very high resolution SEM image of both micro and nano-structures formed within the

surface relief.

Figure 11. (a) Optical image of white light diffraction from large area (1cm2) periodic structuring on stainless steel with 3 parallel spots,

linearly polarized normal to scan direction. Laser exposure was

= 25W, 404kHz with 3ms-1 scan speed, minimising pulse overlap. (b)

SEM image showing clear 1 m pitch periodic structuring at

= 25W average power. The direction of the electric vector is shown in red.

(c) Detailed high resolution SEM image showing nano-structures within the main 1 m pitch surface relief.

5.2 High Power processing at Laser powers at

above 100W.

The results 3 spot (st1 , zero order) multi-beam surface ablation of stainless steel and 1MHz repetition

rate with

= 120W and

= 160W are shown in Figure 12 with increasing numbers of overscans (top

to bottom). The asymmetric pattern, previously observed at

= 25W is still apparent at both

=

120W and

= 160W and structures deepen with increasing exposure, as expected. The 3 spot CGH is

still effective at this power level. With 50 overscans at

= 160W, the surface has oxidised.


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

Figure 12. Optical image of three beam parallel processing with

= 120W/280mms-1 with 1, 5 and 10 overscans,

= 120W/1ms-1 with 2, 10 and 20 overscans, and

= 160W/1ms-1 with 1, 5 and 50 overscans. The assymetric pattern persists. With 50

overscans at

= 160W, the surface has darkened due to oxidation.

Figure 13(a) shows an SEM image of the asymmetric ablation pattern (zero order, st1 ) machined on polished

stainless steel at

= 160W/1MHz/3ms-1 and 1overscan confirming that the cooled SLM diffracts the

radiation into the first orders, even at this extreme exposure, however with low diffraction efficiency as

expected. Figure 13(b) shows an SEM image of the st1 order ablation with 1 m pitch plasmon structuring

observed at the low intensity wing.

Figure 13. Parallel beam surface processing on stainless steel at

= 160W exposure, (a) SEM image of zero and st1 orders with scan

speed 3ms-1, 1 overscan with ~0.5mm offset between scans. First order diffraction efficiency is low. (b) High resolution SEM image of the

overlapped st1 orders showing surface ablation while also displaying periodic plasmon structures at the low intensity wing (RHS)

As diffraction efficiency had reduced significantly at

= 160W, large area parallel beam processing was

carried out at

= 100W/404kHz where phase response was nearly ideal and diffraction efficiency high. A

linear array of four first order rectangular intensity profiles were generated with the Hamamatsu CGH software

and high speed ablation was carried out on stainless steel with a 250mm F-theta lens at 20m/sec scan speed,

figure 14. Pulse overlap was low, fluence F ~0.5Jcm-2 and measured ablation rate was R ~ 4mm3 min-1.


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

Figure 14. Microscope image of four beam (rectangular intensity) high speed large area processing of stainles steel at

=

100W/404kHz/20ms-1 with 400 overscans at fluence F ~ 0.5Jcm-2. Ablation rate R~4mm3min-1. Zero order was neglible.

5.3 More complex phase CGH’s.

Patterns with 11 or more spots were next considered. Figure 15(a) shows multi spots in the form of a letter

M pattern created in the Labview interface. The resulting complex phase CGH, calculated by the Inverse

Fourier Transform (lens and gratings algorithm) is shown in Figure 15(b). The diffracted intensity

distribution with low laser power (

~2W) observed on a screen near the Fourier plane of the first lens

L1 of the 4f system is shown in Figure 15(c). There is reasonable uniformity between the spot intensities

observed here combined with high diffraction efficiency > 90%.

Figure 15. (a) Spot pattern set in Labvew interface representing the letter M, (b) resulting complex phase CGH created by the lens

and gratings algorithm, (c) observed intensity pattern at the Fourier plane of lens L1 of the 4f system showing reasonable spot

uniformity. Zero order spot is in the centre.

Figure 16(a-d) shows a series of letter M patterns drilled on a polished brass coated substrate with the

complex phase CGH of figure 16(b) while increasing the average laser power from

= 90 -250W with

temporal exposures from 10-30ms at 2MHz repetition rate. This corresponds to 20,000 and 60,000 pulse

exposure respectively. While the M pattern is clear at

= 90W and 120W, diffraction efficiency drops

significantly above these powers in accord with reduced phase response. The tails in the spots arise from

the setting of the “laser delay on” in the scanner software which was non optimum, firing the laser a little

too early. As the home position of the scanner was set to (0, 0), the tails rotate due to their positions, micro-

structured around (0, 0).


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

Figure 16. Optical images of surface patterning (letter “M”) on polished brass plate with increasing laser powers, (a)

= 90W,

30ms, (b)

= 120W, 30ms, (c)

= 160W, 10ms, (d)

= 250W, 10ms.. Diffraction efficiency drops above

= 120W as

expected.

5.4 Long term exposure tests

Both SLM devices have been tested for extended periods at high powers with no detrimental effects detected. In

particular, continuous exposure on device-02 for a period of 45minutes at laser power

= 160W and for

several 20minute exposures at

= 220W did not result in any detectable drop in reflected power with time.

No permanent change in SLM phase response characteristics have been detected during this research work.

6. Discussion and Conclusions

Cooled SLMs used as programmable, diffractive optics have tremendous potential for scientific and industrial

applications when handling coherent radiation with high peak and average powers. The research presented in

this paper, we believe, adds significantly to the contributions of Beck [14], Kaakkunen [15,16] and more

recently, Klerks and Eifel [17]. Firstly, the temperature response of two efficiently cooled SLM devices

(Hamamatsu X13139-03) was measured when exposed to continuous, high repetition rate, high average power

1064nm picosecond laser exposure up to

= 220W. These thermal responses (measured using a thermal

camera calibrated for the emissivity of Silicon) were highly linear with gradients of 0.041 C / W (-01) and 0.026

C / W (-02) respectively and are the first, detailed temperature data reported at multi-hundred Watt laser

exposures, higher by a factor of 3.7 (at

= 220W) than that of Klerks and Eifal [17]. The difference in the

thermal characteristics of the two test devices is likely due to a difference in thermal handling design,

proprietary to Hamamatsu. A clear temperature gradient appears across the chip at laser powers above

=

120W exposure, highlighting the efficient axial heat removal. The absolute temperature rise on the best device-

02 with

= 220W was only 5 C at the centre, posing no threat to the viability of the liquid crystal layer or

silicon substrate. Thermal modelling of the SLM structure (with and without water cooling) was carried out

using COMSOL multi-physics and the 3D heat diffusion equation. This model predicted that during laser

exposure, the silicon chip comes to thermal equilibrium only when the SLM was water cooled, reaching 2 C

above ambient compared to the observed 5 C rise with

= 220W exposure (

= 7W absorbed). The

expected chip temperature rises linearly with exposure in accord with observations and in particular, a


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

temperature gradient develops at higher average powers

above 120W. Allowing for photo-absorption in

the organic LC layer, an estimated 0.4C additional temperature rise could be expected. Combining these two

estimates, a 2.5 C rise is within a factor of two of the observed temperature rise.

The thermal response time of our cooled SLMs was approximately 1-2 seconds, observed with the thermal

camera when cutting off the laser power at

= 220W, supporting the very efficient heat removal from the

Silicon chip and consistent with the predicted response time from thermal modelling. As device reflectivity is

97%, then the absorbed power ~ 7W thermal load was removed efficiently at

= 220W incident

laser power.

The critical phase response with increasing exposure on the two cooled devices was measured using

polarisation modulation by the SLM followed by polarisation analysis with quarter waveplate and thin film

polariser (TFP). This unique approach directly generated the phase response curves (for device-01 and -02) up

to

= 140W and

= 200W respectively. Device -02, with the best thermal response, also showed a

superior phase response, achieving full 2 phase change at

= 109W exposure, almost 3 / 2 radians at

= 160W, while at

= 200W, this reduced to just over a phase change. Hence, the operational device

limit is currently at an exposure level of

= 130W. The source of the shifting phase response is likely due to

liquid crystal thickness variations caused by the developing temperature gradient across the chip with increasing

exposure, confirmed by the thermal camera images

The phase response was modelled theoretically, considering both the possible temperature dependence of the

LC birefringence n and thickness of the LC layer L. As the effect of temperature on n appeared neglible, the

phase changes thus appear to be due to variations in the LC layer thickness L with temperature and the

developing temperature gradient across the Silicon chip. This varying thickness affects the phase response so

that, for a given GL, the centre produces a different phase delay (polarisation) than the edges. The beam was

modelled as the sum of three concentric areas, each having a distinct uniform polarization angle and a trial

solution involving the sum of these varying polarisations was compared with the observed response at

=

160W with reasonable agreement. Theory and experiment, as expected, were excellent agreement at lower

average power

= 26W.

At

= 109W exposure, the degree of polarisation remains high with max min max min/ 0.90 P I I I I

while at

= 160W and

= 196W, P 0.75 and 0.64 respectively. A spatial variation of the polarisation state across the reflected beam will occur due to the temperature gradient, hence the reduced value of P. While

Kaakkunen [15] demonstrated cooled SLM exposure with

= 190W continuous wave, no phase or

temperature data were measured during their experiments on laser cutting which was only marginally improved

over the zero order Gaussian beam. Their low diffraction efficiency can be understood from the results achieved

here.

We also presented experimental results of parallel processing on polished metals with continuous powers

from 25W-160W in this paper with interesting results. High speed periodic, 1 m pitch surface micro-structuring

with 25W average power created a continuous, diffractive surface on stainless steel, patterned at a rate of ~

8cm2s-1. Efficient, multi- beam ablation on stainless steel with up to

= 120W was demonstrated while

short (10-30ms) exposures on brass coated stainless steel substrates (11 spot M pattern) showed reduced

diffraction efficiencies above

= 120W.

High power exposure tests were completed with continuous operation at

= 160W for 45 minutes and at

= 220W for several 15minute periods with no degradation observed on device-02. Total exposure times during

these detailed experiments amount to many hours on both devices compared with a 2 hour test at

= 60W by

Klerks and Eifel [17]. If chip damage was going to happen, this would probably happen in seconds, not

minutes, due to thermal runaway at these extreme powers. Any absorbing defect would rapidly heat and boil the

liquid crystal, however, no permanent changes in liquid crystal performance have been detected in this research

work. However, long term changes in diffraction efficiency with high laser exposures could yet be an issue but

would require detailed measurements over much longer periods.


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

One may ask what the damage threshold for irreversible changes in SLM performance will likely be. High

energy pulses (fs-ns) with fluence F > 0.5Jcm-2 might well result in irreversible damage to the structure as

dielectric coating damage thresholds are typically at this level. Damage can also result from very high peak

intensities induced by non-linear (NL) multi-photon absorption, creating absorbing defects in the liquid crystal.

However, as this LCOS technology can handle I > 20GWcm-2, the effect is irrelevant here at peak powers Ppeak

~200MWcm-2. High average powers result in a liquid crystal chip temperature rise through absorption both in

the Silicon, ceramic and copper substrates. As the absorption coefficient of pure silicon at 1064nm is only 10cm-

1 then approximately 63% of the incident radiation would be absorbed in a 1 mm thick Si wafer while the

transmitted radiation will be absorbed in the cooling structure below the Si chip.

In conclusion, we have demonstrated for the first time, temperature and critical phase response of cooled

SLMs exposed to average laser powers exceeding

= 200W at 1064nm, a factor of 3.7 (at

= 220W)

higher than that achieved recently by Klerks and Eifel [17], combined with Ppeak = 200MW peak powers with no

sign of irreversible optical damage. Modelling of the thermal and phase response has also been presented and

compared with experiment. Thermal effects are due mainly to photo-absorption in the Silicon chip and the phase

response limited by LC thickness variations due to the developing temperature gradient across the chip. Parallel

beam micro-structuring up to

= 120W was demonstrated on stainless steel with high diffraction efficiency,

however, higher average powers result in low diffraction efficiency due to the loss of phase range in accord

with expectations. The thermal and optical phase responses are highly correlated, allowing one to understand

device limitations at high exposures. If the temperature gradient could be removed by further improved thermal

cooling, then it may be possible to improve performance above the current operational limit of

= 130W.

Ultimately, testing these devices with both ultrahigh peak intensities exceeding Ppeak = 50GWcm-2 combined

with multi hundred Watt average powers would be highly desirable. Scientific applications such as NL

filamentation in air for control of lightning discharges may be possible [29] where dynamic control of phase,

polarisation and orbital angular momentum of ultrahigh peak power pulses at high repetition rate would be

beneficial. Future expansion of industrial applications in ultra high throughput laser marking, patterning and

machining are likely when high power laser sources are combined with polygon scanners [30] able to achieve

scan speeds on a substrate s > 200ms-1. When combined with a high energy, high average power laser system

and optimised CGHs creating uniform or variable intensity spots [31] massively parallel-beam laser micro-

structuring for industrial applications will be possible.

7. Acknowledgements

We would like to thank Hamamatsu (Japan) for providing two X13139-03 cooled SLM devices for this

collaborative research work and Hamamatsu (Germany) for loan of a 1280 x 1024 SLM controller. We are also

grateful to Dr. Phil Harrison for supplying the polarisation optics and to Professor Miles Padgett of the

University of Glasgow, Scotland for allowing use of the Labview software.

References

1. J. D. McKnight, D. G. Vass and R.M. Sillitto, “Development of a spatial light modulator: a randomly addressed liquid-crystal-over-nMOS array,” Applied Optics, Vol. 28, Issue 22, pp. 4757-4762 (1989)

2 D. J. McKnight, K. M. Johnson and R. A. Serati, “256 × 256 liquid-crystal-on-silicon spatial light modulator”, Applied Optics,

Vol. 33, Issue 14, pp. 2775-2784 (1994) ,

3 M.E. Motamedi, editor, MOEMS (Micro-optical electro-mechanical systems) SPIE Press, Bellingham, Washington (2005)

4 J. Courtial, K. Dholakia, L. Allen, and M. J. Padgett, “Gaussian beams with very high orbital angular momentum,” Opt.

Commun. 144(4-6), 210–213 (1997).

5 N. B. Simpson, K. Dholakia, L. Allen, and M. J. Padgett, “Mechanical equivalence of spin and orbital angular momentum of

light: an optical spanner”, Opt.Lett Vol. 22, No. 1 (1997)


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

https://www.osapublishing.org/ao/issue.cfm?volume=28&issue=22https://www.osapublishing.org/ao/issue.cfm?volume=33&issue=14

6 Y. Hayasaki, T. Sugimoto, A. Takita, N. Nishida, “Variable holographic femtosecond laser processing by use of a spatial light modulator” Applied Physics Letters, 87 p. 031101, (2005)

7 Y. Jin, O. J. Allegre, W. Perrie, K. Abrams, J. Ouyang, E. Fearon, S. P. Edwardson, and G. Dearden, “ Dynamic modulation of spatially structured polarization fields for real-time control of ultrafast laser-material interactions”, Optics Express vol.21, no.21

25333-25343 (2013)

8 Z. Kuang, W. Perrie, D. Liu, S.P. Edwardson and G. Dearden “Ultrafast laser parallel microdrilling using multiple annular beams

generated by a spatial light modulator” J.Phys. D: Applied Physics 47, p.115501 (2014)

9 J. Ouyang, W. Perrie, O,J. Allegre, T. Heil, Y.Jin, E. Fearon, D. Eckford, S.P. Edwardson and G. Dearden “Tailored optical vector fields for ultrashort-pulse laser induced complex surface plasmon structuring” Optics Express vol. 23, no.10 12562-12572 (2015)

10 Jahja I. Trisnadi, Clinton B. Carlisle and Robert Monteverde, "Overview and applications of Grating-Light-Valve-based optical write engines for high-speed digital imaging", Proc. SPIE 5348, MOEMS Display and Imaging Systems II, 52 (January 24,

2004)

11 K. J. Mitchell, S. Turtaev, M. J.Padgett, T. Cizmar and D. J. Phillips, “High speed spatial control of the intensity, phase and

polarisation of vector beams using digital micro-mirror device” Optics Express, 24(25), 29269-29282 (2016)

12 P. Schaaf, editor, “Laser Processing of Materials, Fundamentals, Applications and Developments”, Springer series in Material

science 139 Springer Verlag Heidelberg (2010)

13 J. Leach, G. Sinclair, P. Jordan, J. Courtial, M. J. Padgett, J.Cooper and Z. Laczik, “3D manipulation of particles into crystal

structures using holographic optical tweezers” Optics Express vol.12, 220 (2004).

14 R. J. Beck, J. P. Parry, W. N. MacPherson, A. Waddie, N. J. Weston, J.D. Shephard, and D. P Hand, “Application of cooled spatial light modulator for high power nanosecond laser micromachining” Optics Express, Vol. 18, Issue 16, pp. 17059-17065 (2010)

15 J. J. J. Kaakkunen, I. Vanttaja and P. Laakso, “Fast Micromachining Using Spatial Light Modulator and Galvanometer Scanner with Infrared Pulsed Nanosecond Fiber Laser” JLMN-Journal of Laser Micro/Nanoengineering Vol. 9, No. 1, p37 (2014)

16 J. J. J. Kaakkunen, P. Laakso, and V. Kujanpää, “Adaptive multibeam laser cutting of thin steel sheets with fiber laser using spatial light modulator, Journal of Laser Applications 26, 032008 (2014)

17 T. Klerks and S. Eifel, “Flexible beam shaping system for next generation of process developments in laser micro-machining”, 9th International conference on Photonic technologies, LANE 2016 (Industrial paper)

18 http://www.altechna.com

19 G.R Fowles, “Introduction to Modern Optics” (2nd edition), Dover Publications, New York 1975

20 N. M. Ravindra, B. Sopori, O. H. Gokce, S. X. Cheng, A. Shenoy, L. Jin, S. Aberdrabbo and Y. Chang, “Emissivity measurements and modelling of Silicon-Related Materials:An Overview” International Journal of Thermophysics Vol. 22, No.5,

(Sep.2001)

21 www.crystran.co.uk

22 H.Cabrera, A.Marcano and Y.Castellanos “Absorption coefficients of nearly transparent liquids measured using thermal lens spectrometry”, Condensed Matter Physics Vol. 9, No 2(46), pp. 385–389 (2006)

23 M. Marinelli, H. Zammit, F. Scudieri, S, Martellucci, F. Bloisi and L. Vicari, “Simultaneous Heat Capacity and Thermal-Diffusivity Photoacoastic Measurement at Liquid-Crystal Phase Transitions” Il Nuovo Cimeto, vol.9D, no.7 (1987)

24 F. Peng, H. Cheng, S. Trepathi, R. J. Twieg and S. T. Wu, “Fast-response infrared phase modulator based on polymer network liquid crystal” Opt. Mat. Express vol.5, no.2 p265-273 (2015)

25 B Tan and K Venkatakrishnan, “ A femtosecond laser-induced periodical surface structure on crystalline silicon” J. Micromech. Microeng. (2006) 16 1080

26 F. Garrelie, J. P. Colombier, F. Pigeon, S. Tonchev, N. Faure, M. Bounhalli, S. Reynaud, and O. Parriaux, “Evidence of surface plasmon resonance in ultrafast laser-induced ripples” Optics Express, Vol. 19, Issue 10, pp. 9035-9043, (2011)

27 B. N. Chichkov, C. Momma, S. Nolte, F. von Alvensleben, A. Tünnermann “Femtosecond, picosecond and nanosecond laser ablation of Solids” Appl. Phys. A (1996) 63: 109.

28 H. Wang, X. Liu and Z. M. Zhang, 225 “Absorption Coefficients of Crystalline Silicon at Wavelengths from 500 nm to 1000 nm” Int. J. Thermophys (2013) 34:213-225


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

http://profiles.spiedigitallibrary.org/summary.aspx?DOI=10.1117%2f12.525898&Name=Jahja+I.+Trisnadihttp://profiles.spiedigitallibrary.org/summary.aspx?DOI=10.1117%2f12.525898&Name=Clinton+B.+Carlislehttp://profiles.spiedigitallibrary.org/summary.aspx?DOI=10.1117%2f12.525898&Name=Robert+Monteverdehttp://eprints.gla.ac.uk/view/author/10309.htmlhttp://eprints.gla.ac.uk/view/journal_volume/Optics_Express.htmlhttps://www.osapublishing.org/oe/issue.cfm?volume=18&issue=16http://www.altechna.com/https://www.osapublishing.org/oe/issue.cfm?volume=19&issue=10

29 Jerome Kasparian and Jean-Pierre Wolf “On lightning control using lasers” in Progress in Ultrafast Intense Laser science V, Springer series in Chemical Physics, chapter 6, p109, editors K. Yamanouchi, A. Guilini and K. Ledingham (Heidelberg 2009)

30 B. Jaeggia, B. Neuenschwandera , M. Zimmermann , R. De Loor and L. Penning, “ High throughput ps-laser micro machining with a synchronized polygon line scanner”, 8th International Conference on Photonic Technologies LANE 2014 Industrial Paper.

31 M. Silvennoinen, J. Kaakkunen, K. Paivasaari and P. Vahimaa “Parallel femtosecond laser ablation with individually controlled intensity” Optics Express Vol. 22, No. 3, 2603 (2014)


123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960 A

ccep

ted

Man

uscr

ipt

Documents

University of Manchester - Investigation of the thermal and optical … · 2018. 1. 30. · Investigation of the thermal and optical performance of a Spatial Light Modulator with